University of Birmingham

Mid- to- late Holocene hydroclimatic changes on the Chinese Loess Plateau Sun, Huiling; Bendle, James; Seki, Osamu; Zhou, Aifeng

DOI: 10.1007/s10933-018-0037-9 License: Other (please specify with Rights Statement)

Document Version Peer reviewed version Citation for published version (Harvard): Sun, H, Bendle, J, Seki, O & Zhou, A 2018, 'Mid- to- late Holocene hydroclimatic changes on the Chinese Loess Plateau: evidence from n-alkanes from the sediments of Tianchi Lake', Journal of Paleolimnology, vol. 60, no. 4, pp. 511-523. https://doi.org/10.1007/s10933-018-0037-9

Link to publication on Research at Birmingham portal

Publisher Rights Statement: The final publication is available at Springer via http://doi.org/10.1007/s10933-018-0037-9

General rights Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes permitted by law.

•Users may freely distribute the URL that is used to identify this publication. •Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. •User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?) •Users may not further distribute the material nor use it for the purposes of commercial gain.

Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document.

When citing, please reference the published version. Take down policy While the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has been uploaded in error or has been deemed to be commercially or otherwise sensitive.

If you believe that this is the case for this document, please contact [email protected] providing details and we will remove access to the work immediately and investigate.

Download date: 29. Sep. 2021 Journal of Paleolimnology

Mid- to- late Holocene hydroclimatic changes on the Chinese Loess Plateau: evidence from n-alkanes from the sediments of Tianchi Lake --Manuscript Draft--

Manuscript Number: JOPL-D-17-00035R4 Full Title: Mid- to- late Holocene hydroclimatic changes on the Chinese Loess Plateau: evidence from n-alkanes from the sediments of Tianchi Lake Article Type: Papers Keywords: n-Alkanes · Paq · Lake level · Mid-late Holocene · Loess Plateau Corresponding Author: Aifeng Zhou Lanzhou University Lanzhou, Gansu CHINA Corresponding Author Secondary Information: Corresponding Author's Institution: Lanzhou University Corresponding Author's Secondary Institution: First Author: Huiling Sun First Author Secondary Information: Order of Authors: Huiling Sun James Bendle Osamu Seki Aifeng Zhou Order of Authors Secondary Information:

Funding Information: National Natural Science Foundation of Dr. Huiling Sun China (41761044) National Natural Science Foundation of Dr. Aifeng Zhou China (41771208) China Scholarship Council Dr. Huiling Sun (2009618032)

Abstract: We have reconstructed the history of mid-late Holocene paleohydrological changes in the Chinese Loess Plateau using n-alkane data from a sediment core in Tianchi Lake. We used Paq (the proportion of aquatic macrophytes to the total plant community) to reflect changes in lake water level, with a higher abundance of submerged macrophytes indicating a lower water level and vice versa. The Paq -based hydrological reconstruction agrees with various other lines of evidence, including ACL (average chain length), CPI (carbon preference index), C/N ratio and the n-alkane molecular distribution of the sediments in Tianchi Lake. The results reveal that the lake water level was relatively high during 5.7 to 3.2 ka BP, and decreased gradually thereafter. Our paleohydrological reconstruction is consistent with existing paleoclimate reconstructions from the Loess Plateau, which suggest a humid mid-Holocene, but is asynchronous with paleoclimatic records from central China which indicate an arid mid- Holocene. Overall, our results confirm that the intensity of the rainfall delivered by the EASM (East Asian summer monsoon) is an important factor in affecting paleohydrological changes in the region and can be considered as further evidence for the development of a spatially asynchronous "northern China drought and southern China flood" precipitation pattern during the Holocene. Suggested Reviewers: Cheng Zhao Professor, Nanjing Institute of Geography and Limnology Chinese Academy of

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation Sciences [email protected] Zhonghui Liu Professor, University of Hong Kong [email protected] James Russell associate professor, Brown University [email protected] Response to Reviewers: Dear Professor Whitmore, Thanks you so much for your corrected my manuscript. I accepted all the corrections and “accept all changes”. Best regards! Sincerely, Aifeng

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation Manuscript Click here to download Manuscript Manuscript _2018-03-07 redact.doc Click here to view linked References

1 Mid- to- late Holocene hydroclimatic changes on the Chinese Loess 2 3 Lake 4 Plateau: evidence from n-alkanes from the sediments of Tianchi 5 6 7 8 9 1 Huiling Suna · James Bendleb · Osamu Sekic · Aifeng Zhoud* 10 11 12 2 13 14 3 a. Key Laboratory of Plateau Lake Ecology and Global Change, College of Tourism 15 16 17 4 and Geography, Yunnan Normal University, Kunming, 650500, China 18 19 20 5 b. School of Geography, Earth and Environmental Sciences, University of 21 22 23 6 Birmingham, Edgbaston, Birmingham, B15 2TT, UK 24 25 7 c. Institute of Low Temperature Science, Hokkaido University, N19W8, Kita-ku, 26 27 28 8 Sapporo, 060-0819, Japan 29 30 31 9 d. Key Laboratory of Western China's Environmental Systems（Ministry of Education), 32 33 34 10 College of Earth and Environmental Sciences, Lanzhou University, Lanzhou, 730000, 35 36 11 China 37 38 39 12 40 41 42 13 * Corresponding Author: Aifeng Zhou ([email protected]) 43 44 45 14 Address: 222 Tianshui South Road, Lanzhou, Gansu, P.R.China. 730000 46 47 15 48 Phone: (+86)13893612602 49 50 16 51 52 53 17 Key words 54 55 56 18 n-Alkanes · Paq · Lake level · Mid-late Holocene · Loess Plateau 57 58 59 19 60 61 62 63 64 65 1 20 Abstract 2 3 21 4 5 6 22 We have reconstructed the history of mid-late Holocene paleohydrological changes in 7 8 9 23 the Chinese Loess Plateau using n-alkane data from a sediment core in Tianchi Lake. 10 11 12 24 We used Paq (the proportion of aquatic macrophytes to the total plant community) to 13 14 25 reflect changes in lake water level, with a higher abundance of submerged 15 16 17 26 macrophytes indicating a lower water level and vice versa. The Paq -based 18 19 20 27 hydrological reconstruction agrees with various other lines of evidence, including 21 22 23 28 ACL (average chain length), CPI (carbon preference index), C/N ratio and the 24 25 29 n-alkane molecular distribution of the sediments in Tianchi Lake. The results reveal 26 27 28 30 that the lake water level was relatively high during 5.7 to 3.2 ka BP, and decreased 29 30 31 31 gradually thereafter. Our paleohydrological reconstruction is consistent with existing 32 33 34 32 paleoclimate reconstructions from the Loess Plateau, which suggest a humid 35 36 33 mid-Holocene, but is asynchronous with paleoclimatic records from central China 37 38 39 34 which indicate an arid mid-Holocene. Overall, our results confirm that the intensity of 40 41 42 35 the rainfall delivered by the EASM (East Asian summer monsoon) is an important 43 44 45 36 factor in affecting paleohydrological changes in the region and can be considered as 46 47 37 48 further evidence for the development of a spatially asynchronous “northern China 49 50 38 drought and southern China flood” precipitation pattern during the Holocene. 51 52 53 39 54 55 56 40 57 58 41 59 60 61 62 63 64 65 1 42 Introduction 2 3 43 4 5 6 44 Climatic and environmental changes in the Chinese Loess Plateau are mainly 7 8 9 45 controlled by the EASM, which directly affects almost all aspects of the hydrology 10 11 12 46 and ecology of East Asia (Clift and Plumb 2008). An increase in EASM intensity 13 14 47 would be expected to result in a northward movement of the rainfall belt in China and 15 16 17 48 a corresponding rainfall increase in the Loess Plateau (Chen et al. 2008). Many 18 19 20 49 regional paleoclimatic records have been produced from this semi-arid, monsoon 21 22 23 50 marginal zone (Zhao et al. 2010; Dong et al. 2012; Liu and Feng 2012; Lu et al. 2013; 24 25 51 Qiang et al. 2013). However, regional high-resolution paleohydrological 26 27 28 52 reconstructions are extremely limited because proxies or archives that record ancient 29 30 31 53 hydrological conditions, with good age control, are scarce on the Loess Plateau. A 32 33 34 54 humid mid-Holocene has been proposed based on a pollen-based record (Chen et al. 35 36 55 2015a) and a hydrogen isotope reconstruction of long-chain n-alkanes (Rao et al. 37 38 39 56 2016) from Gonghai Lake, one of the few natural lakes on the Loess Plateau. Their 40 41 42 57 paleohydrological reconstruction is inconsistent with records from the core 43 44 45 58 monsoon-controlled regions of central China. It shows an arid interval from 7.0-3.0 ka 46 47 59 48 BP (Xie et al. 2013; Zhu et al. 2017). Therefore, more high-resolution lacustrine 49 50 60 reconstructions of hydroclimatic variations during the mid-late Holocene are needed 51 52 53 61 to explore the underlying mechanism of this asynchronous hydroclimatic variability. 54 55 56 62 Here, a high-resolution lacustrine record based on n-alkanes of sediments from 57 58 63 59 Tianchi Lake on the Loess Plateau will be discussed. 60 61 62 63 64 65 1 64 n-Alkanes preserved in lake sediments can be used to infer variations in the 2 3 65 composition and origin of organic inputs to the lacustrine environment, because they 4 5 6 66 are widely preserved in various environmental contexts, such as plants, soils and 7 8 9 67 lacustrine sediments, and can resist degradation actions (Meyers 1997). In particular 10 11 12 68 the Average Chain Length (ACL) (Poynter and Eglinton 1990), Carbon Preference 13 14 69 Index (CPI) (Meyers and Ishiwatari 1993), and P (Ficken et al. 2000) n-alkane 15 aq 16 17 70 indices, have been widely used in paleoenvironmental research (Nichols et al. 2006; 18 19 20 71 He et al. 2014). In general, terrestrial plants and emergent macrophytes are typically 21 22 23 72 dominated by the long-chain length homologues (C27-C33) (Ficken et al. 2000; Gao et 24 25 73 al. 2011), while submerged and floating-leaved macrophytes mainly produce C and 26 23 27 28 74 C25 n-alkanes (Ficken et al. 2000), and short chain ones are produced by algae and 29 30 31 75 bacteria (Cranwell et al. 1987). Consequently, higher ACL and CPI are commonly 32 33 34 76 considered to be predominantly produced by terrestrial plants. A higher Paq may result 35 36 77 from an increase in submerged macrophytes in combination with a recession of the 37 38 39 78 terrestrial plants around the lake. Moreover, the biomass of submerged macrophytes is 40 41 42 79 related to the variation of the water table (Wagner and Falter 2002; Liu et al. 2015), 43 44 45 80 and lake level fluctuations have the potential to simultaneously constrain the spatial 46 47 81 48 distribution and the biomass of submerged macrophytes in a lake (Duarte and Kalf 49 50 82 1986; Hudon 1997; Middelboe and Markager 1997). Howerver, Aichner et al. (2010) 51 52 53 83 and Liu et al. (2015) found that higher amounts of long chain n-alkanes can be 54 55 56 84 produced by submerged macrophytes in several lakes. Therefore, it is necessary to 57 58 85 59 understand the extent to which long chain n-alkanes in lacustrine sediments are 60 61 62 63 64 65 1 86 influenced by terrestrial plants and submerged macrophytes in a study lake when 2 3 87 reconstructing the paleoenvironments. 4 5 6 88 In this study, we first define the potential sources and the contributions from the 7 8 9 89 various plants (e.g. terrigenous plants vs. submerged macrophytes) in Tianchi Lake. 10 11 12 90 Second, we give an interpretation of the proxies (Paq, ACL, CPI of n-alkanes, and 13 14 91 C/N), especially P as an effective indicator of lake level changes in Tianchi Lake. 15 aq 16 17 92 Additionally, we seek to compare regional climate reconstructions with those from 18 19 20 93 Tianchi Lake and other nearby sites to confirm a spatially asynchronous 21 22 23 94 hydroclimatic variability occurred in China during the Holocene. 24 25 95 26 27 28 96 Study site 29 30 31 97 32 33 34 98 Tianchi Lake (lat. 35º15’55’’N, long. 106º18’43’’E, elevation 2430 m a.s.l.) is a small 35 36 99 freshwater alpine lake located in the Liupan Mountains, southwestern Loess Plateau, 37 38 39 100 northwest China (Fig. 1a). The length of the lake from east to west is 250 m and the 40 41 42 101 width is 120 m. The maximum water depth is 8.2 m, and the lake covers an area of 43 44 4 2 45 102 2×10 m (Fig. 1b). The lake receives no surface run off, and it is fed by meteoric 46 47 103 48 water and groundwater recharge. There is no apparent surface outflow, except for a 49 50 104 possible transient outflow in the western part of the lake basin, which is possibly 51 52 53 105 active during the rainy season. The mean annual temperature is 8.2 °C and mean 54 55 56 106 annual precipitation is 677 mm based on data from the nearest meteorological station 57 58 107 59 (Liupan Mountain station, at 2845 m a.s.l.). Most of the precipitation occurs as 60 61 62 63 64 65 1 108 rainfall during summer, accounting for nearly 72.2% of the annual total. The 2 3 109 vegetation of the upland slopes of the lake is dominated by shrubs and steppe. Grassy 4 5 6 110 steppe with sparse shrub covers the north slopes, and shrubs dominate the south 7 8 9 111 slopes (Zhao et al. 2010). Emergent (Phragmites australis (Cav.) Trin. ex Steud) and 10 11 12 112 submerged (Potamogeton sp. and Chara sp.) macrophytes are widely distributed in 13 14 113 the shallow areas of the lake (Fig. 1d), but floating-leaved macrophytes are absent, 15 16 17 114 based on our field observations in 2010. 18 19 20 115 21 22 23 116 Materials and methods 24 25 117 26 27 28 118 Field sampling 29 30 31 119 32 33 34 120 Two parallel sediment cores of lengths 11.2 m (GSA07-1) and 10.4 m (GSB07-1) 35 36 121 were collected using a UWITEC piston corer system (6 cm in diameter) from the lake 37 38 39 122 center in 2007 (Fig. 1b). The lithology of core GSB07-1 consisted of alternating 40 41 42 123 brown-colored sandy clay and grey-brownish clay between 1040-746 cm, and 43 44 45 124 grey-brownish clay above 746 cm. The sediments are characterized by 1 to 46 47 125 48 2-mm-thick organic detritus-rich laminations (Fig. 1c), which yielded abundant 49 50 126 terrestrial macrofossils for radiocarbon dating. Fifty-six down-core sedimentary 51 52 53 127 samples were taken at 15-cm intervals throughout core GSB07-1. Additionally, we 54 55 56 128 collected six surface soil samples, three surface lake sediments, nine dominant 57 58 129 59 terrestrial plant samples (Cedrus sp., Larix sp., Abies sp., Betula sp., Rosa sp., Rubus 60 61 62 63 64 65 1 130 sp., Salix sp., Berberis sp. and Artemisia sp.), which surround the lake, and one 2 3 131 emergent macrophyte (P. australis) and two submerged macrophytes (Potamogeton 4 5 6 132 and Chara) within the lake for modern process study. All the above samples were 7 8 9 133 carried out for TOC, TN analyses, and lipid extraction. 10 11 12 134 13 14 135 Laboratory analyses 15 16 17 136 18 19 20 137 Samples for TOC and TN measurements were pretreated with 10 ml of 10% HCl to 21 22 23 138 remove carbonates, washed with distilled water until the pH was neutral, and then 24 25 139 measured using a CE Model 440 Elemental Analyzer. The C/N ratio was derived from 26 27 28 140 the ratio of TOC and TN. n-Alkanes were extracted based on methods described 29 30 31 141 previously (Kawamura et al. 2003) in G-MOL lab of the University of Glasgow. 32 33 34 142 Briefly, 2-10 g of freeze-dried, homogenized sediment were transferred to a test tube 35 36 143 and hydrolyzed with 15 ml of 0.3 M KOH dissolved in 95:5 37 38 39 144 methanol/dichloromethane-extracted water. The samples were then hydrolyzed and 40 41 42 145 centrifuged and the supernatant and pipetted into a round-bottomed flask. The 43 44 45 146 sediment was then extracted three times with 10 ml dichloromethane/methanol (3:1) 46 47 147 48 using ultrasonication. The extracts were combined and concentrated, using a rotary 49 50 148 evaporator, under vacuum and then separated into neutral and acidic fractions using 51 52 53 149 the methods of Kawamura (1995). The neutral fraction was further separated using 54 55 56 150 silica gel column chromatography to get n-alkane fraction. Dried n-alkane fraction 57 58 151 59 was redissolved in hexane and analyzed using a gas chromatograph (GC; Shimadzu 60 61 62 63 64 65 1 152 2010) with a flame ionization detector (FID) and hydrogen as carrier gas at constant 2 3 153 pressure (190 kPa). Separation of the different compounds was achieved using an 4 5 6 154 identical column (length: 60 m, diameter: 0.25 mm, film thickness: 0.25 μm, coating: 7 8 9 155 100 % dimethyl-polysiloxane). The gas chromatograph temperature program was set 10 11 -1 -1 12 156 to increase from 50 -120 °C at 30 °C min , then 120 -310 °C at 5 °C min , with a 13 14 157 final isothermal time of 20 min at 300 °C. Compound identification was confirmed by 15 16 17 158 GC/MS (Shimadzu OP2010-Plus Mass Spectrometer (MS) interfaced with a 18 19 20 159 Shimadzu 2010 GC) based on retention times and mass spectra. 21 22 23 160 The n-alkane proxies (equation (1) from Poynter and Eglinton (1990); equation (2) 24 25 161 from Marzi et al. (1993); and equation (3) from Ficken et al. (2000) were calculated 26 27 28 162 as follows: 29 30 31 163 ACL=(19*C19+20*C20+21*C21+⋯+33*C33)/(C19+C20+C21+⋯+C33) (1) 32 33 34 164 CPI =7/8*(C19+C21+C23+⋯+C33)/(C20+C22+C24+⋯+C32) (2) 35 36 165 P = (C +C )/(C +C +C +C ) (3) 37 aq 23 25 23 25 29 31 38 39 166 where Ci is the concentration of n-alkane of i number of carbon. 40 41 42 167 43 44 45 168 Age model 46 47 169 48 49 50 170 The chronology of core GSA07-1 used in this study mainly consists of 19 dates from 51 52 53 171 Zhao et al. (2010) and 6 new dates (Table 1). All 14C dates were measured in the AMS 54 55 56 172 Dating Laboratory of Beijing University and are based on the leaves of terrestrial 57 58 173 59 plants. The ages were calibrated to calendar years before present (AD 1950) using the 60 61 62 63 64 65 1 174 program CALIB Rev. 5.0.1 with the IntCal04 calibration data set (Reimer et al. 2004). 2 3 175 The depths of characteristic laminations in cores GSA07-1 and core GSB07-1 are 4 5 6 176 consistent. Therefore, the chronology of core GSB07-1 was calibrated based on the 7 8 9 177 corresponding depths in GSA07-1 (Table 1). The chronology indicates that the age of 10 11 12 178 core GSB07-1 spans the past 5720 years (Fig. 2). The average accumulation rate 13 14 179 based on the age-depth model is about 1.85 mm a-1. 15 16 17 180 18 19 20 181 Results 21 22 23 182 24 25 183 n-Alkane distributions and Paq variations in modern vegetation 26 184 27 185 28 29 30 186 The Paq index has been proposed as an indicator of the relative contributions of 31 32 187 n-alkanes from submerged/floating aquatic plants versus those from emergent and 33 34 35 188 terrestrial plants in the lake. Generally, Paq <0.1 corresponds to terrestrial plants, 36 37 38 189 0.1-0.4 to emergent macrophytes, and 0.4-1.0 to submerged/floating macrophytes 39 40 41 190 (Ficken et al. 2000). In this study, average Paq values and n-alkane molecular 42 43 191 distribution patterns vary considerably in the three types of plant material (terrestrial, 44 45 46 192 and emergent and submerged macrophytes: Fig. 3a-c). Terrestrial plants (Fig. 3a), 47 48 49 193 which have a lower average Paq value (0.18), are dominated by the n-C31 homologue. 50 51 52 194 Emergent macrophytes (Fig. 3b) growing in the near-shore environment are mainly 53 54 195 dominated by the n-C27 homologue and have a higher Paq value (0.65). In contrast, 55 56 57 196 n-C23 is the dominant homologue in the submerged macrophytes (Fig. 3c) with a 58 59 60 197 secondary peak at n-C25. The average Paq value of submerged macrophytes is 0.93. In 61 62 63 64 65 1 198 addition, a bimodal n-alkane distribution pattern with high abundances at n-C23 and 2 3 199 n-C is observed in the surface sediments of Tianchi Lake which have an average P 4 31 aq 5 6 200 value of 0.51 (Fig. 3d), indicating a specific mixture of inputs from terrestrial plants 7 8 9 201 and submerged macrophytes. The distribution pattern for the surface soil has an 10 11 12 202 overwhelming preponderance of the n-C31 homologue, and the average Paq value of 13 14 203 the surface soils is 0.25 (Fig. 3e). 15 16 17 204 18 19 20 205 n-Alkane proxies and C/N ratios in the down-core sediments 21 22 23 206 24 25 207 Time series of the various sedimentary parameters are illustrated in Fig. 4. The 26 27 28 208 records span the last 5.7 ka BP. Paq (Fig. 4a) ranges from 0.32 to 0.78 with a mean of 29 30 31 209 0.56. The average Paq value is 0.46 during 5.7-3.2 ka BP, and 0.65 from 3.2 ka BP to 32 33 34 210 the present. It is noteworthy that prior 3.2 ka BP most of the Paq values are less than 35 36 211 0.52, while subsequently they are greater than 0.52. ACL ranges from 25 to 29 with a 37 38 39 212 mean of 27.4 (Fig. 4b). The CPI values range from 1.7 to 8.8 with a mean of 5.2 over 40 41 42 213 the last 5.7 ka BP (Fig. 4c). The C/N ratios (Fig. 4d) range from 9.5 to 26 with a mean 43 44 45 214 of 15. The ACL, CPI, and C/N ratios exhibit similar patterns of variation, and they all 46 47 215 48 exhibit an obvious shift at 3.2 ka BP, as do the Paq values. The threshold values of 49 50 216 ACL, CPI, and C/N ratios are almost the same as their average values. 51 52 53 217 54 55 56 218 Discussion 57 58 219 59 60 61 62 63 64 65 1 220 Sources of organic matter to the lake 2 3 221 4 5 6 222 The organic component of lake sediments represents a pool of organic matter derived 7 8 9 223 from the decomposing detritus of aquatic plants growing in the littoral and marginal 10 11 12 224 zone of the lake and from terrestrial plants growing in the catchment (Meyers and 13 14 225 Ishiwatari 1993; Meyers 1997). Lacustrine sediment n-alkanes often have multiple 15 16 17 226 sources, including terrestrial plants, aquatic macrophytes and lower organisms. 18 19 20 227 Generally, n-alkane distributions of terrestrial and emergent plants tend to exhibit 21 22 23 228 high proportions of the n-C31 homologue (Rielley et al. 1991; Ficken et al. 2000; 24 25 229 Sachse et al. 2006), whereas those of submerged and floating plants are generally 26 27 28 230 dominated by n-C23 and n-C25 homologues (Ficken et al. 2000; Gao et al. 2011; Seki 29 30 31 231 et al. 2012). Therefore, n-alkanes can be used to identify local and regional sources of 32 33 34 232 organic matter. However, recent studies have indicated that aquatic plants also make a 35 36 233 large contribution to the long chain n-alkanes in lake sediments (Aichner et al. 2010; 37 38 39 234 Liu et al. 2015; Liu and Liu 2016). For example, Liu et al. (2015) found that the long 40 41 42 235 chain n-alkanes produced by submerged plants in Qinghai Lake had a significant 43 44 45 236 influence on n-C27 and n-C29 alkanes in sediments. Even for the same submerged 46 47 237 48 plant (Potamogeton sp.) from 16 Tibetan Plateau lakes, the distribution patterns of all 49 50 238 the n-alkane homologs show obvious differences (Liu and Liu 2016). It is thus 51 52 53 239 necessary to make a distinction between the various sources that contribute to the 54 55 56 240 organic matter in given study area. At present, the two types of submerged 57 58 241 59 macrophytes (Potamogeton and Chara) in Tianchi Lake are dominated by mid-chain 60 61 62 63 64 65 1 242 n-alkanes (n-C23 and n-C25) (Fig. 3c). The average Paq value is as high as 0.93. There 2 3 243 is no evidence that they exhibit relatively high abundance of long chain n-alkanes as 4 5 6 244 Liu and Liu (2016) described. The unimodal distribution pattern with the maxima at 7 8 9 245 n-C31 alkanes and relatively low Paq values of modern terrestrial plants (Fig. 3a) and 10 11 12 246 surface soils (Fig. 3e) in Tianchi Lake, suggest again that n-C31 alkanes can be traced 13 14 247 to terrestrial plant inputs and not to lake macrophytes. In addition, a bimodal 15 16 17 248 molecular distribution pattern with major peaks at the n-C23 and n-C31 homologues in 18 19 20 249 the surface lake sediments (Fig. 3d) probably represents a combination of inputs from 21 22 23 250 submerged macrophytes (Fig. 3c) and terrestrial plants (Fig. 3a)/emergent (Fig. 3b). 24 25 251 Our observations are consistent with those of previous studies (Cranwell 1984; Ficken 26 27 28 252 et al. 2000; Gao et al. 2011; Street et al. 2013), which indicate that Paq can be used to 29 30 31 253 reflect the contribution from submerged macrophytes. 32 33 34 254 35 36 255 Interpretation of n-alkane indices 37 38 39 256 40 41 42 257 It has been demonstrated that the abundance of submerged macrophytes in lake is 43 44 45 258 affected by irradiance and the littoral slope (Hudon 1997; Hudon et al. 2000; 46 47 259 48 Cheruvelil and Soranno 2008). Thus, lake level has the potential to constrain the 49 50 260 spatial distribution of submerged macrophytes via both a reduction in light intensity 51 52 53 261 (Duarte and Kalf 1986; Middelboe and Markager 1997) and a change in the spatial 54 55 56 262 extent of the littoral habitat (Hudon 1997). Hence, changes in the relative inputs of 57 58 263 59 submerged macrophytes can potentially be ascribed to fluctuations in lake level. Most 60 61 62 63 64 65 1 264 of the submerged macrophytes in Tianchi Lake are distributed in a shallow area close 2 3 265 to the shoreline and very few floating macrophytes can be observed (Fig. 1d). 4 5 6 266 Potamogeton and Chara are the two dominant submerged species, which grow in a 7 8 9 267 narrow zone down to a depth of ~1.2 m. A bathymetric survey of Tianchi Lake (Fig. 10 11 12 268 1b) reveals that the shoreline forms a narrow shelf from a depth of 0.5 m down to 2.8 13 14 269 m, followed by a steep slope that causes a decrease in the occurrence of submerged 15 16 17 270 macrophytes. Assuming that the basic bathymetry of the basin has remained similar 18 19 20 271 through time, the reductions in lake level would result in a relatively larger shelf area, 21 22 23 272 which would produce an expansion of the shallow-water habitat for submerged 24 25 273 macrophytes. Accordingly, the lower P values in Tianchi Lake could be interpreted 26 aq 27 28 274 as reflecting less abundance of submerged macrophytes and the raising of lake level. 29 30 31 275 On the other hand, the contribution from terrestrial plants can also exert an influence 32 33 34 276 on Paq values since the Paq index is a proxy for evaluating the contribution of 35 36 277 n-alkanes from submerged/floating aquatic plants relative to emergent and terrestrial 37 38 39 278 plants (Ficken et al. 2000). During intervals of high rainfall and lake level, an 40 41 42 279 increased contribution of terrestrial plant material delivered by increased catchment 43 44 45 280 rain and runoff could lower Paq values, and vice versa. This coincides with the 46 47 281 48 findings of Liu and Liu (2016), which indicated a negative relationship between the 49 50 282 Paq value of surface lake sediments and the water level of Qinghai Lake. 51 52 53 283 As an important parameter of n-alkanes, the climate implications of ACL have 54 55 56 284 been discussed a lot in the literature, but there is no unified agreement as to their 57 58 285 59 interpretation because ACL often appears highly specific to regional or local 60 61 62 63 64 65 1 286 conditions (Ling et al. 2017). Furthermore, ACL can not be used to reconstruct 2 3 287 temperature or precipitation change if the plant species or sedimentary environment in 4 5 6 288 the catchment area underwent considerable change (in parallel with or forced by 7 8 9 289 climatic variation) (Pu et al. 2010). CPI is another n-alkane index, which has been 10 11 12 290 widely accepted as an indicator for terrestrial sources of sedimentary organic matter. 13 14 291 Terrestrial plants have abundant long-chain n-alkanes, and show distinct odd-even 15 16 17 292 predominance, thus their CPI is always greater than 5. On the contrary, CPI values of 18 19 20 293 the aquatic plants and planktonic bacteria are considerably lower than those usually 21 22 23 294 reported for terrestrial plant sourced n-alkanes (Cranwell 1987). 24 25 295 26 27 28 296 Reconstruction of the lake-level evolution 29 30 31 297 32 33 34 298 n-Alkane based records and C/N ratios from Tianchi Lake are presented in Fig. 4. 35 36 299 Overall, there is an obvious shift at ~3.2 ka BP among the various proxies. During 37 38 39 300 5.7-3.2 ka BP, the ACL (Fig. 4b) and CPI (Fig. 4c) proxies show relatively high 40 41 42 301 values in the core. CPI values almost greater than 5 and ACL values range from 27 to 43 44 45 302 29, likely indicate a predominance of terrestrial plant inputs to the lake basin. The 46 47 303 48 results are supported by the higher C/N ratios (mean >15, with occasional values up to 49 50 304 26) (Fig. 4d) in this phase since the C/N ratios from terrestrial plants and emergent 51 52 53 305 macrophytes can be as high as 20 (Lamb et al. 2004). On the other hand, relatively 54 55 56 306 lower Paq values (0.32 - 0.56) (Fig. 4a) and a contrary changing trend of Paq with ACL, 57 58 307 59 CPI, C/N ratios, suggest a recession of the submerged macrophytes growing in 60 61 62 63 64 65 1 308 Tianchi Lake. Based on interpretations discussed above, less abundance of submerged 2 3 309 macrophytes input and more abundance of terrestrial plants input to the sediments are 4 5 6 310 likely a response to relatively high lake levels during this phase in Tianchi Lake. 7 8 9 311 We interpret the increase in average Paq values (0.51-0.78) and a decrease in ACL 10 11 12 312 (25-27) and CPI (1.6-5.4) after 3.2 ka BP (Fig. 4a) as corresponding to an increase in 13 14 313 the proportion of submerged and floating-leaved macrophytes, and a decrease in 15 16 17 314 terrestrial inputs. Furthermore, the C/N ratios are generally low (<15) during this 18 19 20 315 interval. In view of the absence of floating-leaved macrophytes in Tianchi Lake, 21 22 23 316 based on our field observations, the increasing contribution from submerged 24 25 317 macrophytes accordingly indicates a gradually falling lake level from 3.2 ka BP. 26 27 28 318 The variations in lake level inferred by the n-alkanes record is also supported by a 29 30 31 319 shift in pollen assemblages from Tianchi Lake, which indicate that closed canopy 32 33 34 320 forest was replaced by an open landscape at around 3.0 ka BP (Zhao et al. 2010). 35 36 321 Another high-resolution pollen record from the Dadiwan peatland (Fig. 1a), 50 km 37 38 39 322 southwest of Tianchi Lake on the Loess Plateau, also reveals a significant decrease in 40 41 42 323 tree pollen frequencies at around 3.0 ka BP (An et al. 2003). 43 44 45 324 46 47 325 48 Asynchronous hydroclimatic variability 49 50 326 51 52 53 327 The Paq record from Tianchi Lake reveals a transition from higher lake levels to lower 54 55 56 328 lake levels after 3.2 ka BP, and thus wetter conditions during 5.7-3.2 ka BP and drier 57 58 329 59 conditions after 3.2 ka BP (Fig. 5d). This accords with other paleoclimatic records 60 61 62 63 64 65 1 330 from the nearby Chinese Loess Plateau (Lu et al. 2013; Chen et al. 2015a; Liu et al. 2 3 331 2015; Rao et al. 2016). The pollen-based annual precipitation reconstruction from 4 5 6 332 Tianchi Lake suggests a rapid precipitation decrease since ~3.3 ka BP (Fig. 5e; Chen 7 8 9 333 et al. 2015a). Another pollen-based annual precipitation reconstruction from nearby 10 11 12 334 Gonghai Lake (Fig. 5f) reveals a humid interval around 8-3 ka BP (Chen et al. 2015a). 13 14 335 A recent study of palaeosol development as an indicator of the strength of the EASM 15 16 17 336 (Wang et al. 2014) suggests a wet interval during 8.6-3.2 ka BP in the Chinese Loess 18 19 20 337 Plateau. In addition, a TOC record from the Dadiwan peat profile also revealed a 21 22 23 338 similar pattern of wet and dry episodes as at Tianchi Lake (Zhou et al. 1996; Huang et 24 25 339 al. 2013). This evidence supports the contention that a moist climate was a 26 27 28 340 widespread phenomenon on the Chinese Loess Plateau during the mid-Holocene. It is 29 30 31 341 in accord with the gradually decreasing solar insolation (Fig. 5g). However, the 32 33 34 342 paleohydrological conditions reconstructed from Dajiuhu peatland (Fig. 5c) in the 35 36 343 middle reaches of the Yangtze River of central China are in contrast with these 37 38 39 344 previous paleoclimatic records. Changes in the aerobic bacteria-derived hopanoid flux 40 41 42 345 in Dajiuhu peatland (Fig. 5c) imply relatively arid conditions from 7.0-3.0 ka BP and 43 44 45 346 relatively wet conditions from 3.0-1.0 ka BP (Xie et al. 2013; Huang et al. 2013; He et 46 47 347 48 al. 2015). Another late-Holocene paleohydrological reconstruction based on sediment 49 50 348 grain-size and n-alkane data from Longgan Lake in the middle and lower reaches of 51 52 53 349 the Yangtze River (Xue et al. 2017), indicated drought conditions from 4 to 2.7 ka BP 54 55 18 56 350 and a humid interval from 2.7 to 1.2 ka BP. In addition, the studies on the δ O (Fig. 57 58 351 59 5a) and the flux of soil-derived magnetic minerals preserved (Fig. 5b) in stalagmite 60 61 62 63 64 65 1 352 HS4 from Heshang cave in central China also revealed a relatively arid interval from 2 3 353 6.7 to 3.4 ka BP (Hu et al. 2008; Zhu et al. 2017). Therefore, it seems that mid-late 4 5 6 354 Holocene paleohydrological evolution was asynchronous in the middle reaches of the 7 8 9 355 Yangtze River of central China and in the Yellow River region of north China. 10 11 12 356 Tianchi Lake is located in the ‘far-field’ northwestern marginal region of the 13 14 357 EASM, whereas the Dajiuhu peatland (Fig. 1a) is located in the ‘core’ monsoonal area 15 16 17 358 of the EASM (Qian et al. 2007). Summer rainfall is the predominant contributor to the 18 19 20 359 annual precipitation at both sites (Gao and Xie 2014). The northwards advance of the 21 22 23 360 rainfall front resulting from an enhanced EASM intensity could result in increased 24 25 361 precipitation in the marginal region of the EASM but decreased precipitation in the 26 27 28 362 core monsoonal area of EASM (Ding et al. 2008; Rao et al. 2016). The occurrence of 29 30 31 363 this contrasting spatial pattern of moisture conditions, with more frequent droughts in 32 33 34 364 north China and more frequent floods in the mid-low Yangtze River valley during 35 36 365 summer, has also been observed during the last few decades (Gemmer et al. 2004; 37 38 39 366 Qian and Lin 2005; Zhai et al. 2005). It has been designated the “northern China 40 41 42 367 drought and southern China flood” precipitation pattern (Zhou et al. 2009), and is also 43 44 45 368 evident on millennial and centennial time scales (Chen et al. 2015b). 46 47 369 48 Previous workers have analyzed the main factors responsible for the 49 50 370 asynchronous pattern of hydroclimatic variability between the marginal and core 51 52 53 371 monsoonal area of EASM in China. For example, He et al. (2014) suggested that 54 55 56 372 terrestrial temperature-induced evaporation changes and the extent of the Asian 57 58 373 59 monsoonal front could potentially explain the out-of-phase pattern of hydrological 60 61 62 63 64 65 1 374 changes during the mid-Holocene. Chen et al. (2015a) emphasized that insolation 2 3 375 forcing, especially the tropical ocean conditions might be responsible for the abrupt 4 5 6 376 decline at 3.3 ka. Chen et al. (2015b) concluded that ENSO is one of the most 7 8 9 377 important factors affecting the precipitation of monsoonal northern and central China 10 11 12 378 on the centennial scale. Rao et al. (2016) emphasized the important influence of the 13 14 379 west-east thermal gradient in the equatorial Pacific on the climate of monsoonal China. 15 16 17 380 Zhu et al. (2017) concluded that a mid-Holocene reduction in ENSO intensity was 18 19 20 381 related to a decrease in storm frequency in the middle reaches of Yangtze River 21 22 23 382 between 6.7 and 3.4 ka BP. Finally, it is likely that the sea surface temperature (SST) 24 25 383 anomaly in the equatorial Pacific during the mid-Holocene probably played a key role 26 27 28 384 in facilitating the influence of ENSO on the asynchronous pattern of precipitation in 29 30 31 385 the marginal and core monsoonal area of the EASM in China. 32 33 34 386 35 36 387 37 Conclusions 38 39 388 40 41 42 389 We have used the record of n-alkanes extracted from a lacustrine sediment core from 43 44 45 390 Tianchi Lake on the Chinese Loess Plateau to reconstruct lake-level variations during 46 47 391 48 the past 5.7 ka BP. Paq values and C/N ratios through the sequence in general exhibit a 49 50 392 gradually increasing trend through the past 5.7 ka BP, indicating an increasing (and 51 52 53 393 more variable) abundance of submerged macrophytes in response to a falling lake 54 55 56 394 level. Terrestrial plants dominated the record before 3.2 ka BP, and subsequently 57 58 395 59 there was a shift to the dominance of submerged macrophytes. The predominance of 60 61 62 63 64 65 1 396 terrestrial plants agree with higher ACL, higher CPI, and lower Paq values from 2 3 397 5.7-3.2 ka BP, whereas the dominance of submerged macrophytes resulted in lower 4 5 6 398 ACL, lower CPI, and higher Paq values after 3.2 ka BP. These changes indicate a 7 8 9 399 relatively humid interval during 5.7-3.2 ka BP and a drier but more variable interval 10 11 12 400 after 3.2 ka BP on the Chinese Loess Plateau. These findings are consistent with 13 14 401 previous paleoclimatic reconstructions for the Loess Plateau which indicate a humid 15 16 17 402 mid-Holocene. However, they are in disagreement with paleoclimatic records from 18 19 20 403 central China, which indicate an arid mid-Holocene. Overall, this spatial pattern 21 22 23 404 indicates that an enhanced intensity of monsoon rainfall delivered by the EASM 24 25 405 during the mid-Holocene was an important factor in affecting paleohydrological 26 27 28 406 changes in the region. 29 30 31 407 32 33 34 408 Acknowledgments 35 36 409 37 38 39 410 We thank Dr.Christopher Gallacher and Dr. Heiko Moossen for their training and 40 41 42 411 help with laboratory analyses. This research was supported by grants from the 43 44 45 412 National Science Foundation of China (NSFC Grants 41761044 and 41771208). We 46 47 413 48 thank the China Scholarship Council (CSC) for funding a 20-month visit (File no. 49 50 414 2009618032) by Huiling Sun to work with Dr. James Bendle (now at the University 51 52 53 415 of Birmingham) as a joint Ph.D. student (Lanzhou-Glasgow) at the G-MOL 54 55 56 416 laboratory in Glasgow. 57 58 417 59 60 61 62 63 64 65 1 418 References 2 3 419 4 5 420 An CB, Feng ZD, Tang LY (2003) Evidence of a humid mid-Holocene in the western part of 6 7 421 8 Chinese Loess Plateau. Chin Sci Bull 48: 2472-2479 9 10 422 Chen FH, Yu ZC, Yang ML, Ito E, Wang SM, Madsen DB, Huang XZ, Zhao Y, Sato T, Birks 11 12 13 423 HJB, Boomer I, Chen JH, An CB, Wuennemann B (2008) Holocene moisture evolution in 14 15 16 424 arid central Asia and its out-of-phase relationship with Asian monsoon history. Quat Sci Rev 17 18 425 19 27: 351-364 20 21 426 Chen FH, Xu QH, Chen JH, Birks HJB, Liu JB, Zhang SR, Jin LY, An CB, Telford RJ, Cao 22 23 24 427 XY, Wang ZL, Zhang XJ, Selvaraj K, Lu HY, Li YC, Zheng Z, Wang HP, Zhou AF, Dong 25 26 27 428 GH, Zhang JW, Huang XZ, Bloemendal J, Rao ZG (2015a) East Asian summer monsoon 28 29 30 429 precipitation variability since the last deglaciation. Sci Rep 5: 11186 31 32 430 Chen JH, Chen FH, Feng S, Huang W, Liu JB, Zhou AF (2015b) Hydroclimatic changes in 33 34 35 431 China and surroundings during the Medieval Climate Anomaly and Little Ice Age: spatial 36 37 38 432 patterns and possible mechanisms. Quat Sci Rev 107: 98-111 39 40 41 433 Cheruvelil KS, Soranno PA (2008) Relationships between lake macrophyte cover and lake 42 43 434 and landscape features. Aquat Bot 88: 219-227 44 45 46 435 Clift PD, Plumb RA (2008) The Asian monsoon: causes, history and effects. Cambridge 47 48 49 436 University Press, Cambridge 50 51 52 437 Cranwell PA (1984) Lipid geochemistry of sediments from Upton Broad, a small productive 53 54 438 lake. Org Geochem 7: 25-37 55 56 57 439 Ding YH, Wang ZY, Sun Y (2008) Inter-decadal variation of the summer precipitation in 58 59 60 440 East China and its association with decreasing Asian summer monsoon. Part I: Observed 61 62 63 64 65 1 441 evidences. Int J Climatol 28: 1139-1161 2 3 442 Dong GH, Yang Y, Zhao Y, Zhou AF, Zhang XJ, Li XB, Chen FH (2012) Human settlement 4 5 6 443 and human-environment interactions during the historical period in Zhuanglang County, 7 8 9 444 western Loess Plateau, China. Quat Int 281: 78-83 10 11 12 445 Duarte CM, Kalf J (1986) Littoral slope as a predictor of the maximum biomass of submerged 13 14 446 macrophyte communities. Limnol Oceanogr 31: 1072-1080 15 16 17 447 Ficken KJ, Li B, Swain DL, Eglinton G (2000) An n-alkane proxy for the sedimentary input 18 19 20 448 of submerged/floating freshwater aquatic macrophytes. Org Geochem 31: 745-749 21 22 23 449 Gao L, Hou JZ, Toney J, MacDonald DL, Huang YS (2011) Mathematical modeling of the 24 25 450 aquatic macrophyte inputs of mid-chain n-alkyl lipids to lake sediments: implications for 26 27 28 451 interpreting compound specific hydrogen isotopic records. Geochim Cosmochim Acta 75: 29 30 31 452 3781-3791 32 33 34 453 Gao T, Xie LA (2014) Study on progress of the trends and physical causes of extreme 35 36 454 precipitation in China during the last 50 years. Adv Earth Sci 29: 577-589 37 38 39 455 Gemmer M, Becker S, Jiang T (2004) Observed monthly precipitation trends in China 40 41 42 456 1951-2002. Theor Appl Clim 77: 39-45 43 44 45 457 He YX, Zheng YW, Pan AD, Zhao C, Sun YY, Song M, Zheng Z, Liu ZH (2014) 46 47 458 Biomarker-based reconstructions of Holocene lake-level changes at Lake Gahai on the 48 49 50 459 northeastern Tibetan Plateau. The Holocene 24: 405-412 51 52 53 460 He YX, Zhao C, Zheng Z, Liu ZH, Wang N, Li J, Cheddadi R (2015) Peatland evolution and 54 55 56 461 associated environmental changes in central China over the past 40,000 years. Quat Res 84: 57 58 462 255-261 59 60 61 62 63 64 65 1 463 Huang XY, Xue JT, Wang XX, Meyers PA, Huang JH, Xie SC (2013) Paleoclimate influence 2 3 464 on early diagenesis of plant triterpenes in the Dajiuhu peatland, central China. Geochim 4 5 6 465 Cosmochim Acta 123: 106-119 7 8 9 466 Hudon C (1997) Impact of water level fluctuations on St. Lawrence River aquatic vegetation. 10 11 12 467 Can J Fish Aquat Sci 54: 2853-2865 13 14 468 Hudon C, Lalonde S, Gagnon P (2000) Ranking the effects of site exposure, plant growth 15 16 17 469 form, water depth, and transparency on aquatic plant biomass. Can J Fish Aquat Sci 57: 31-42 18 19 20 470 Kawamura K (1995) Land-derived lipid class compounds in the deep-sea sediments and 21 22 23 471 marine aerosols from North Pacific. In: Biogeochemical Processes and Ocean Flux in the 24 25 472 Western Pacific. Terra Scientific Publishing Company, Tokyo, pp 31-51 26 27 28 473 Kawamura K, Ishimura Y, Yamazaki K (2003) Four years' observations of terrestrial lipid 29 30 31 474 class compounds in marine aerosols from the western North Pacific. Global Biogeochem Cy 32 33 34 475 17: 3-1-3-19 35 36 476 Lamb AL, Leng MJ, Mohammed MU, Lamb HF (2004) Holocene climate and vegetation 37 38 39 477 change in the Main Ethiopian Rift Valley, inferred from the composition (C/N and δ13C) of 40 41 42 478 lacustrine organic matter. Quat Sci Rev 23: 881-891 43 44 45 479 Ling Y, Zheng MP, Sun Q, Dai XQ (2017) Last deglacial climatic variability in Tibetan 46 47 480 Plateau as inferred from n-alkanes in a sediment core from Lake Zabuye. Quat Int 454: 15-24 48 49 50 481 Liu FG, Feng ZD (2012) A dramatic climatic transition at ~4,000 cal. yr BP and its cultural 51 52 53 482 responses in Chinese cultural domains. The Holocene 22: 1181-1197 54 55 56 483 Liu H, Liu WG (2016) n-Alkane distributions and concentrations in algae, submerged plants 57 58 484 and terrestrial plants from the Qinghai-Tibetan Plateau. Org Geochem 99: 10-22 59 60 61 62 63 64 65 1 485 Liu JB, Chen JH, Zhang XJ, Li Y, Rao ZG, Chen FH (2015) Holocene East Asian summer 2 3 486 monsoon records in northern China and their inconsistency with Chinese stalagmite δ18O 4 5 6 487 records. Earth-Sci Rev 148: 194-208 7 8 9 488 Liu WG, Yang H, Wang HY, An ZS, Wang Z, Leng Q (2015) Carbon isotope composition of 10 11 12 489 long chain leaf wax n-alkanes in lake sediments: A dual indicator of paleoenvironment in the 13 14 490 Qinghai-Tibet Plateau. Org Geochem 83: 190-201 15 16 17 491 Lu HY, Yi SW, Liu ZY, Mason JA, Jiang DB, Cheng J, Stevens T, Xu ZW, Zhang EL, Jin 18 19 20 492 LY (2013) Variation of East Asian monsoon precipitation during the past 21 ky and potential 21 22 23 493 CO2 forcing. Geology 41: 1023-1026 24 25 494 Marzi R, Torkelson BE, Olson RK (1993) A revised carbon preference index. Org Geochem 26 27 28 495 20: 1303-1306 29 30 31 496 Meyers PA (1997) Organic geochemical proxies of paleoceanographic, paleolimnologic, and 32 33 34 497 paleoclimatic processes. Org Geochem 27: 213-250 35 36 498 Meyers PA, Ishiwatari R (1993) Lacustrine organic geochemistry - an overview of indicators 37 38 39 499 of organic matter sources and diagenesis in lake sediments. Org Geochem 20: 867-900 40 41 42 500 Middelboe AL, Markager S (1997) Depth limits and minimum light requirements of 43 44 45 501 freshwater macrophytes. Freshw Biol 37: 553-568 46 47 502 Nichols JE, Booth RK, Jackson ST, Pendall EG, Huang YS (2006) Paleohydrologic 48 49 50 503 reconstruction based on n-alkane distributions in ombrotrophic peat. Org Geochem 37: 51 52 53 504 1505-1513 54 55 56 505 Poynter J, Eglinton G (1990) Molecular composition of three sediments from hole 717c: The 57 58 506 Bengal fan. In: Proceedings of the Ocean Drilling Program: Scientific results, 116: pp. 59 60 61 62 63 64 65 1 507 155-161 2 3 508 Pu, Y, Zhang HC, Lei GL, Chang FQ, Yang MS, Zhang WX, Lei YB, Yang LQ, Pang YZ 4 5 6 509 (2010) Climate variability recorded by n-alkanes of paleolake sediment in Qaidam Basin on 7 8 9 510 the northeast Tibetan Plateau in late MIS3. China Earth Sci: 53: 624-631 (In Chinese) 10 11 12 511 Qian W, Lin X (2005) Regional trends in recent precipitation indices in China. Meteorol 13 14 512 Atmos Phys 90: 193-207 15 16 17 513 Qian WH, Lin X, Zhu YF, Xu Y, Fu JL (2007) Climatic regime shift and decadal anomalous 18 19 20 514 events in China. Clim Change 84: 167-189 21 22 23 515 Qiang MR, Song L, Chen FH, Li MZ, Liu XX, Wang Q (2013) A 16 ka lake-level record 24 25 516 inferred from macrofossils in a sediment core from Genggahai Lake, northeastern 26 27 28 517 Qinghai-Tibetan Plateau (China). J Paleolimnol 49: 575-590 29 30 31 518 Rao ZG, Jia GD, Li YX, Chen JH, Xu QH, Chen FH (2016) Asynchronous evolution of the 32 33 34 519 isotopic composition and amount of precipitation in north China during the Holocene 35 36 520 revealed by a record of compound-specific carbon and hydrogen isotopes of long-chain 37 38 39 521 n-alkanes from an alpine lake. Earth Planet Sci Lett 446: 68-76 40 41 42 522 Reimer PJ, Baillie MGL, Bard E, Bayliss A, Beck JW, Bertrand CJH, Blackwell PG, Buck 43 44 45 523 CE, Burr GS, Cutler KB, Damon PE, Edwards RL, Fairbanks RG, Friedrich M, Guilderson 46 47 524 TP, Hogg AG, Hughen KA, Kromer B, McCormac FG, Manning SW, Ramsey CB, Reimer 48 49 50 525 RW, Remmele S, Southon JR, Stuiver M, Talamo S, Taylor FW, van der Plicht, J, 51 52 53 526 Weyhenmeyer CE, (2004) IntCal04 terrestrial radiocarbon age calibration, 0-26 cal kyr BP. 54 55 56 527 Radiocarbon 46: 1029-1058 57 58 528 Rielley G, Collier RJ, Jones DM, Eglinton G (1991) The biogeochemistry of Ellesmere Lake, 59 60 61 62 63 64 65 1 529 UK-I: source correlation of leaf wax inputs to the sedimentary lipid record. Org Geochem 17: 2 3 530 901-912 4 5 6 531 Sachse D, Radke J, Gleixner G (2006) δD values of individual n-alkanes from terrestrial 7 8 9 532 plants along a climatic gradient - Implications for the sedimentary biomarker record. Org 10 11 12 533 Geochem 37: 469-483 13 14 534 Seki O, Harada N, Sato M, Kawamura K, Ijiri A, Nakatsuka T (2012) Assessment for 15 16 17 535 paleoclimatic utility of terrestrial biomarker records in the Okhotsk Sea sediments. Deep Sea 18 19 20 536 Research Part II: Topical Studies in Oceanography 61: 85-92 21 22 23 537 Street JH, Anderson RS, Rosenbauer RJ, Paytan A (2013) n-Alkane evidence for the onset of 24 25 538 wetter conditions in the Sierra Nevada, California (USA) at the mid-late Holocene transition, 26 27 28 539 ~ 3.0 ka. Quat Res 79: 14-23 29 30 31 540 Vogts A, Moossen H, Rommerskirchen F, Rullkötter J (2009) Distribution patterns and stable 32 33 34 541 carbon isotopic composition of alkanes and alkan-1-ols from plant waxes of African rain 35 36 542 forest and savanna C species. Org Geochem 40: 1037-1054 37 3 38 39 543 Wagner T, Falter CM (2002) Response of an aquatic macrophyte community to fluctuating 40 41 42 544 water levels in an oligotrophic lake. Lake Reservoir Manag 18: 52-65 43 44 45 545 Wang HP, Chen JH, Zhang XJ, Chen FH (2014) Palaeosol development in the Chinese Loess 46 47 546 Plateau as an indicator of the strength of the East Asian summer monsoon: Evidence for a 48 49 50 547 mid-Holocene maximum. Quat Int 334: 155-164 51 52 53 548 Wang SW, Huang JB, Wen XY, Zhu JH (2008) Evidence and modeling study of droughts in 54 55 56 549 China during 4-2 ka BP. Chin Sci Bull 53: 2215-2221 57 58 550 Xie SC, Evershed RP, Huang XY, Zhu ZM, Pancost RD, Meyers PA, Gong LF, Hu CY, 59 60 61 62 63 64 65 1 551 Huang JH, Zhang SH, Gu YS, Zhu JY (2013) Concordant monsoon-driven postglacial 2 3 552 hydrological changes in peat and stalagmite records and their impacts on prehistoric cultures 4 5 6 553 in central China. Geology 41: 827-830 7 8 9 554 Xue JT, Li JJ, Dang XY, Meyers PA, Huang XY (2017) Paleohydrological changes over the 10 11 12 555 last 4,000 years in the middle and lower reaches of the Yangtze River: Evidence from particle 13 14 556 size and n-alkanes from Longgan Lake. The Holocene 27: 1318-1324 15 16 17 557 Zhai PM, Zhang XB, Wan H, Pan XH (2005) Trends in total precipitation and frequency of 18 19 20 558 daily precipitation extremes over China. J Climate 18: 1096-1108 21 22 23 559 Zhao Y, Chen FH, Zhou AF, Yu ZC, Zhang K (2010) Vegetation history, climate change and 24 25 560 human activities over the last 6200 years on the Liupan Mountains in the southwestern Loess 26 27 28 561 Plateau in central China. Palaeogeogr Palaeoclimatol Palaeoecol 293: 197-205 29 30 31 562 Zhou TJ, Gong DY, Li J, Li B (2009) Detecting and understanding the multi-decadal 32 33 34 563 variability of the East Asian Summer Monsoon-recent progress and state of affairs. Meteorol 35 36 564 Z 18: 455-467 37 38 39 565 Zhou WJ, Douglas JD, Stephen CP, Timothy AJ, Li XQ, Minze S, An ZS, Eiji M, Dong GR 40 41 42 566 (1996) Variability of monsoon climate in East Asia at the end of the last glaciation. Quat Res 43 44 45 567 46: 219-229 46 47 568 Zhu ZM, Feinberg JM, Xie SC, Bourne MD, Huang CJ, Hu CY, Cheng H (2017) Holocene 48 49 50 569 ENSO-related cyclic storms recorded by magnetic minerals in speleothems of central China. 51 52 53 570 Proc Natl Acad Sci 114: 852-857 54 55 56 571 57 58 59 572 Figure captions 60 61 62 63 64 65 1 573 Fig. 1. (a) Location of Tianchi Lake in North China. Solid dots represent the study 2 3 574 area and other study sites referenced in the text. The map shows the correlation 4 5 6 575 coefficients between summer precipitation in China and summer monsoon intensity 7 8 9 576 from 1951-2000 (Wang et al. 2008), (b) schematic representation of the bathymetry of 10 11 12 577 Tianchi Lake (depths in m), (c) laminated structure of the sediment cores from 13 14 578 Tianchi Lake, (d) photo of submerged macrophytes in the shallow area of Tianchi 15 16 17 579 Lake 18 19 20 580 21 22 23 581 Fig. 2. Age-depth model for core GSB07-1 from Tianchi Lake 24 25 582 26 27 28 583 Fig. 3. Histogram of the molecular distributions of n-alkanes from (a) modern 29 30 31 584 terrestrial plants, (b) modern emergent macrophytes, (c) modern submerged 32 33 34 585 macrophytes, (d) surface lake sediments and (e) surface soils from around Tianchi 35 36 586 Lake. Only odd carbon number distributions are shown for the n-alkanes 37 38 39 587 40 41 42 588 Fig. 4. Time series of sedimentary parameters for core GSB07-1 from Tianchi Lake 43 44 45 589 over the past 5.7 ka BP. (a) Paq values based-on n-alkanes, (b) n-alkane ACL, (c) 46 47 590 48 n-alkane CPI, (d) C/N ratios 49 50 591 51 52 53 592 Fig. 5. Comparison of regional paleohydrological records. (a) Heshang cave 54 55 18 56 593 speleothem δ O records (Hu et al. 2008), (b) the flux of soil-derived magnetic 57 58 594 59 minerals (IRMsoft-flux) preserved in stalagmite HS4 (Zhu et al. 2017), (c) hopanoids 60 61 62 63 64 65 1 595 flux in Dajiuhu peatland (Xie et al. 2013), (d) Paq values based-on n-alkanes in 2 3 596 Tianchi Lake, (e) pollen-based reconstruction of mean annual precipitation (MAP) 4 5 6 597 from Tianchi Lake (Chen et al. 2015a), (f) pollen-based reconstruction of mean annual 7 8 9 598 precipitation (MAP) from Gonghai Lake (Chen et al. 2015a), (g) July Insolation at 10 11 12 599 30 °N (Berger and Loutre 1991) 13 14 600 15 16 17 601 18 19 20 602 21 22 23 603 24 25 604 26 27 28 605 29 30 31 606 32 33 34 607 35 36 608 37 38 39 609 40 41 42 610 43 44 45 611 46 47 612 48 49 50 613 51 52 53 614 54 55 56 615 57 58 59 616 60 61 62 63 64 65 1 617 Table 1. AMS radiocarbon dates of core GSA07-1 in Tianchi Lake and the chronology 2 3 618 of core GSB07-1 based on depth calibration with core GSA07-1 4 5 6 619 7 8 9 Core GSA07-1 Core GSB07-1 10 Depth Material δ13C 14C date Error Calibrated age Calibrated Calibrated age 11 (cm) dated (‰ VPDB) (yr BP) (±yr) (Cal yr BP-2σ range) depth (cm) (Cal yr BP) 12 162 Tree leaves -29.0 680 30 619±56 174 662±21 13 14 183 Tree leaves -26.9 855 35 740±47 193 776±27 15 221 Tree leaves -21.1 1080 35 963±35 228 1009±33 16 17 260 Tree leaves -11.6 1255 30 1169±42 240 1088±23 18 302 Tree leaves -18.7 1440 45 1378±37 257 1192±14 19 383 Tree leaves -21.0 1775 30 1793±30 366 1768±58 20 21 436 Tree leaves -24.3 2060 30 2089±42 443 2217±82 22 489 Tree leaves -23.1 2355 30 2398±52 509 2633±37 23 510 Tree leaves -17.6 2520 35 2537±49 526 2719±26 24 25 518 Tree leaves -12.7 2585 40 2580±52 532 2754±25 26 554 Tree leaves -26.3 2415 35 2789±51 545 2830±25 27 28 570 Tree leaves -25.4 2830 30 2898±45 554 2891±30 29 600 Tree leaves -16.6 2895 45 3097±51 593 3130±43 30 660 Tree leaves -19.7 3300 30 3520±43 658 3538±38 31 32 701 Tree leaves -30.3 3400 30 3793±36 700 3810±19 33 732 Tree leaves -24.4 3720 35 4002±33 730 4012±28 34 35 751 Tree leaves -20.2 3935 35 4149±45 754 4174±21 36 815 Tree leaves -14.7 4100 40 4547±33 816 4540±29 37 848 Tree leaves -19.5 4230 35 4726±33 838 4653±28 38 39 893 Tree leaves -18.3 4400 40 4937±41 882 4899±33 40 966 Tree leaves -16.3 4495 40 5277±33 957 5290±25 41 1014 Tree leaves -25.6 4820 40 5513±32 992 5497±26 42 43 1049 Tree leaves -27.1 4970 35 5689±34 1023 5673±33 44 1086 Tree leaves -25.4 5030 35 5870±33 45 46 1114 Tree leaves -24.4 5440 70 6038±41 47 620 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 fig1 Click here to download Figure Fig.1-2016.8.5.tif fig2 Click here to download Figure Fig.2-GSB-new.pdf 0 200 400 600 Depth (cm) 800 1000

0 1000 2000 3000 4000 5000 6000 Age (Cal yr BP) Figure3 Click here to download Figure Fig.3.2018-03-08.pdf 5 0 T e r r e s t r i a l ( n = 9 ) ( a ) 4 0 P = 0 . 1 8 a q

3 0

2 0

1 0

0 ( b ) 5 0 E m e r g e n t ( n = 1 ) 4 0 P = 0 . 6 5 a q

3 0

2 0

1 0

) 0

% ( c ) ( 5 0 S u b m e r g e d ( n = 2 ) e c 4 0 P = 0 . 9 3 n a q a d

n 3 0

u b

a 2 0

e v i

t 1 0 a l

e 0 R 5 0 ( d ) S u r f a c e S e d i m e n t s ( n = 3 ) 4 0 P = 0 . 5 1 a q 3 0

2 0

1 0

0 ( e ) 5 0 S u r f a c e S o i l ( n = 6 ) P = 0 . 2 5 4 0 a q

3 0

2 0

1 0

0 C 2 1 C 2 3 C 2 5 C 2 7 C 2 9 C 3 1 C 3 3

C a r b o n n u m b e r o f n - a l k a n e s

Figure4 Click here to download Figure Fig.4..pdf 0 1 0 0 0 2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 0 6 0 0 0

( d ) 2 5

2 0 C / N 1 5

1 0 1 0 ( c ) 8

I 6 P

C 4

2 2 9 ( b ) 2 8 A C L

2 7

2 6 ( a ) 0 . 8

0 . 6 q a P 0 . 4

0 . 2 0 1 0 0 0 2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 0 6 0 0 0 A g e ( C a l y r B P ) Figure5 Click here to download Figure Fig.5.2018-03-08.tif table1 Click here to download Table Table1.docx

Table 1

Core GSA07-1 Core GSB07-1

Depth Material δ13C 14C date Error Calibrated age Calibrated Calibrated age (cm) dated (‰ VPDB) (yr BP) (±yr) (Cal yr BP-2σ range) depth (cm) (Cal yr BP) 162 Tree leaves -29.0 680 30 619±56 174 662±21 183 Tree leaves -26.9 855 35 740±47 193 776±27 221 Tree leaves -21.1 1080 35 963±35 228 1009±33 260 Tree leaves -11.6 1255 30 1169±42 240 1088±23 302 Tree leaves -18.7 1440 45 1378±37 257 1192±14 383 Tree leaves -21.0 1775 30 1793±30 366 1768±58 436 Tree leaves -24.3 2060 30 2089±42 443 2217±82 489 Tree leaves -23.1 2355 30 2398±52 509 2633±37 510 Tree leaves -17.6 2520 35 2537±49 526 2719±26 518 Tree leaves -12.7 2585 40 2580±52 532 2754±25 554 Tree leaves -26.3 2415 35 2789±51 545 2830±25 570 Tree leaves -25.4 2830 30 2898±45 554 2891±30 600 Tree leaves -16.6 2895 45 3097±51 593 3130±43 660 Tree leaves -19.7 3300 30 3520±43 658 3538±38 701 Tree leaves -30.3 3400 30 3793±36 700 3810±19 732 Tree leaves -24.4 3720 35 4002±33 730 4012±28 751 Tree leaves -20.2 3935 35 4149±45 754 4174±21 815 Tree leaves -14.7 4100 40 4547±33 816 4540±29 848 Tree leaves -19.5 4230 35 4726±33 838 4653±28 893 Tree leaves -18.3 4400 40 4937±41 882 4899±33 966 Tree leaves -16.3 4495 40 5277±33 957 5290±25 1014 Tree leaves -25.6 4820 40 5513±32 992 5497±26 1049 Tree leaves -27.1 4970 35 5689±34 1023 5673±33 1086 Tree leaves -25.4 5030 35 5870±33 1060 5871±35 1114 Tree leaves -24.4 5440 70 6038±41 1094 6078±58